1. Field of the Invention
This invention relates to a wafer flatness evaluation method, a wafer flatness evaluation apparatus which carries out the evaluation method, a wafer manufacturing method using the evaluation method, a wafer quality assurance method using the evaluation method, a semiconductor device manufacturing method using the evaluation method and a semiconductor device manufacturing method using a wafer evaluated by the evaluation method.
2. Description of the Related Art
The depth of focus gets closer to the limit as the pattern size of a semiconductor device is more miniaturized.
In the 100 nm generation, it is important to suppress a variation in the focus which is a CD (Critical Dimension: the size of a pattern transferred on a resist) variation factor as far as possible and reconsider each focus variation factor. As the focus variation factor, an influence by the wafer flatness is significant and it is required to further increase the flatness of a wafer.
In the conventional manufacturing method of a wafer represented by a silicon wafer, for example, a thin disk is cut out by slicing a single crystal ingot and sequentially subjected to various steps such as beveling, lapping, etching steps. Then, at least a main surface (one surface or both surfaces) of the wafer is polished to make a mirror-like surface and a mirror finished wafer (or PW) is formed. Elements are formed on the mirror finished wafer by use of an exposure device or the like to manufacture a device.
Wafers used in the device process are not limited to the mirror finished wafer and an epitaxial wafer having an epitaxial layer formed on a mirror finished wafer and an anneal wafer subjected to the heat treatment can be used in some cases. Further, a wafer having an added value attached to the mirror finished wafer, for example, an SOI wafer formed by laminating two mirror finished wafers with an oxide film disposed. therebetween can be used in some cases (the above various types of wafers are generally referred to as wafers).
The wafers are formed by setting working processes (working process condition) to attain wafer quality corresponding to a specification of flatness or the like set in the device process based on the specification.
As shown in FIG. 32, as the conventional definition of the flatness of the wafer, SFQR (Site flatness quality requirements=total range of wafer topography relative to focal plane) which is derived based on the wafer flatness as viewed from the thickness distribution of the wafer, that is, the thickness distribution obtained when it is assumed that the wafer is completely chucked on an ideal plane is widely used. In FIG. 32, a reference symbol 3201 indicates the cross sectional shape of a site in the free standing state and a reference symbol 3202 indicates the cross sectional shape obtained when the site 3201 is completely chucked on the ideal plane.
The flatness required in the photolithography is flatness (which is referred to as “SFQRSR” in this specification) which an exposure device, for example, a scanner senses in a state where the wafer is chucked on a wafer holder like the case of actual exposure as shown in FIGS. 33C and 34C from the viewpoint of a focus budget.
FIGS. 33A to 33C are cross sectional views showing the cross sectional shapes of a site (which is referred to as a “full site” in this specification) lying near the center of the wafer.
FIG. 33A shows the cross sectional shape of the full site in the free standing state and FIG. 33B shows the cross sectional shape when the full site is completely chucked on the ideal plane and the flatness SFQR thereof. Further, FIG. 33C shows the cross sectional shape when the full site is chucked on a pin-chuck type wafer holder and the flatness SFQRSR which the scanner senses.
When the pin-chuck type wafer holder is used, it is reported in a document 1 that the flatness SFQR of the full site becomes substantially the same as the flatness SFQRSR which the scanner senses.
FIGS. 34A to 34C are cross sectional views showing the cross sectional shapes of a site (which is referred to as a “partial site” in this specification) lying near the peripheral edge portion of the wafer.
FIG. 34A shows the cross sectional shape of the partial site in the free standing state and FIG. 34B shows the cross sectional shape when the partial site is completely chucked on the ideal plane and the flatness SFQR thereof. Further, FIG. 34C shows the cross sectional shape when the partial site is chucked on a pin-chuck type wafer holder and the flatness SFQRSR which the scanner senses.
As is understood by comparing FIGS. 34B and 34C, a large influence by the interaction between the edge shape of the wafer, the stage shape of the wafer holder and the structure of the chucking portion appears on the flatness of the partial site. Therefore, the flatness SFQRSR which the scanner senses becomes largely different from the result of the flatness SFQR when the shape of the wafer edge is derived and the site is completely chucked on the ideal plane. For example, it is reported in a document 2 that the flatness of the partial site is different from the flatness SFQR depending on the location of the wafer in which a hold groove in the outermost peripheral portion of the stage of the wafer holder is formed.
Further, in a document 3, the relation between the polishing pressure used at the CMP time and the distance from the wafer center, the wafer flatness after chucking and the edge roll-off is described.
Document 1: T. Fujisawa et al, “Analysis of Wafer Flatness for CD Control in Photolithography”, Proc. SPIE 4691, pp. 802-809, 2002
Document 2: N. Poduje, “Edge Effect on Flatness for 130 nm Technology and beyond”, Proc. SEMI Japan Silicon Wafer Workshop, pp. 101-106, 2001
Document 3: Tetsuo Fukuda, “JEITA Flatness Study IV and The Impact of Edge Roll-off on CMP”, [ONLINE] 2002.04.17<, Advanced Wafer Geometry Task Force 2002 SEMICON/Europe (Munich), [Retrieved on Sep. 3, 2002], Internet<hyperlink symbology omitted>
As described above, in the conventional wafer flatness provided when the wafer is purchased, the flatness of a holder used at the exposure time and the interaction between the wafer and the holder are not taken into consideration. Therefore, particularly, the flatness of the partial site becomes different from the flatness which the scanner, that is, exposure device senses. As a result, a wafer which is evaluated to have high flatness by use of the conventional evaluation standard, for example, evaluation standard by SFQR cannot exhibit an expected performance when the wafer is actually chucked on the holder of the exposure device. Then, a focus variation amount exceeds the budget and it becomes difficult to suppress a CD variation within a sufficiently permissible range.